机器视觉中几种颜色优势的图像分割算法比较与研究

摘要

为了获得实用、可靠的植物识别方法，必须在识别植物过程中进行图像分割。本文针对具有明显颜色优势的图像，采用Otsu阀值算法及边缘检测分割算法，对图像进行处理，并对其结果进行对比分析和优劣比较。得出结论：对具有明显颜色优势的图像，采用边缘检测分割算法的应用效果较好；对比边缘检测算法中常用的Prewitt算子、Roberts算子和Sobel算子的输出效果，认为Sobel算子不仅能输出较为理想的效果，同时还可大幅度减少运算时间。%　　In order to obtain a practical, reliable plant identification method, image segmentation must be in the process of identifying plants. Advantage for the obvious color image segmentation algorithm of Otsu threshold algorithm and edge detection, image processing, and compare their results were analyzed and the pros and cons. Conclusion: the advantage of obvious color image, edge detection segmentation algorithm is better; contrast edge detection algorithm used Prewitt operator, Roberts operator output effect of the Son, and Sobel operator, Sobel operator not only can output the desired results, but can also significantly reduce the computation time.